(Informal) Chat with Dr. Earl Geddes of GedLee Audio

[Newman]

What do you get out of a "sweep" with respect to peak playback capability?

I only used the word “sweep” because you introduced it, when answering my thoughts on Erin’s test. I hope we are talking here about Erin’s test? I therefore direct your question to Erin. @hardisj

 

[amarsicola]

I might be profane in the room, but i've always wondered why distortion in Olive's index does not matter much, maybe because at Harman they went for simple THD and they didn't consider the Gedlee metric?
If i read the papers from Geddes (e.g. The Perception of Distortion) his metric achieves a .96 correlation with subjective preference scores, quite an impressive result, comparable with Olive's index.
Any explanation from senior members?

 

[Newman]

I might be profane in the room, but i've always wondered why distortion in Olive's index does not matter much, maybe because at Harman they went for simple THD and they didn't consider the Gedlee metric?
If i read the papers from Geddes (e.g. The Perception of Distortion) his metric achieves a .96 correlation with subjective preference scores, quite an impressive result, comparable with Olive's index.
Any explanation from senior members?

Because the patterns and levels of distortion found in modern hifi gear don't correspond to the distortions that the Geddes investigation found to be most sensitive to audibility.

re Harman -- I assume you are talking about speakers -- Geddes himself says speaker drivers simply don't generate distortion patterns that cause audibility problems.

I vaguely recall raising your question with Dr Geddes on a forum many years ago, and IIRC he wasn't that fussed that his metric wasn't 'taken up' across the industry, because, although it's an important bit of intelligence to understand the importance of accounting for masking in deciding whether a distortion pattern and level is audible or preferable, he said that modern electronics and speaker drivers are not generating distortion patterns and levels that are audible issues. I emphasize that I am going from memory here, and hope I recall correctly.

cheers

 

[headshake]

Very little fluff to that vid!

I could listen to a couple more hours of you both talking.

Maybe Dr. Earl can find the time to do this every few months? Build up some topics and then crush them with facts.

@hardisj You mentioned seeing dips 1-2khz in some speakers. Do you think the head shadow masks some of the errors in this range? To my newbie mind, it kinda makes sense that many XO's are in this range. Our ears expect dips here.

"A sample of frequency response of ears: green curve: left ear XL(f) blue curve: right ear XR(f) for a sound source from upward front. "

https://en.wikipedia.org/wiki/Head-related_transfer_function